Factors Impacting Consumers’ Purchase Intention of Electric Vehicles in China: Based on the Integration of Theory of Planned Behaviour and Norm Activation Model

Abstract

Understanding the factors that drive consumers to purchase electric vehicles (EVs) is critical to achieving decarbonization of China’s transportation sector, as well as mitigating global warming. This study aims to construct a research model based on altruistic and self-interested perspectives by integrating the Theory of Planned Behaviour (TPB) and Norm Activation Model (NAM) to predict the psychological factors that influence Chinese consumers’ intention to purchase EVs. Data were collected from 867 participants in China and empirically tested using Structural Equation Modeling (SEM). Self-interested factors, namely subjective norms, attitudes and perceived behavioural control, all had a significant positive effect on EV purchase intention. Additionally, the results showed that personal norms had the greatest effect on EV purchase intention. It was also found that awareness of consequence, ascription of responsibility and subjective norms were positive predictors of personal norms. Awareness of consequence had a positive effect on both the ascription of responsibility and attitudes. The findings contribute to understanding the psychological drivers of Chinese consumers’ intention to purchase EVs and can provide decision-making references for policy makers and manufacturers.

1. Introduction

In recent years, global climate change has caused a wide range of environmental impacts and threats, such as sea level rise, increased extreme weather events and ecosystem destruction. Greenhouse gas (GHG) emissions are the primary cause of global climate change, and fossil fuel consumption in industry and transportation is considered to be the main factor to the increase in GHG emissions [1,2,3]. The global transportation sector produces more than 8.4 billion tonnes of CO₂ emissions, which is about 22.83% of the total global carbon emissions [4]. Private vehicles account for about 50% of the energy use and carbon emissions from the transportation sector; thus, decarbonizing transportation is significant in mitigating global climate change [5]. EVs are powered by electricity and can achieve zero emissions during the use phase. Meanwhile, compared with vehicles relying on traditional fuels, EVs can reduce CO₂ emissions by 30–50% and improve fuel efficiency by 40–60% [6], which means the promotion of EVs has become a major way for the transportation sector to achieve carbon reduction [7]. Therefore, it is crucial to understand the stimuli associated with EV purchase decisions [8].

The transportation sector is a high-energy-consuming and high-polluting industry and is one of the major sources of carbon emissions in China [9]. The Chinese Government attaches great importance to decarbonisation and the promotion of EVs in the transport sector. The 14th Five-Year Plan for the Development of a Modern Comprehensive Transport System states that “by 2025, China’s comprehensive transport greening level should achieve a substantial breakthrough, and the transport industry should be comprehensively promoted in the direction of green and low-carbon transformation [10].” In November 2020, the New Energy Vehicle Industry Development Plan (2021–2035), issued by the General Office of the State Council, proposed that by 2035, pure EVs would become the mainstream of new vehicles for sale [11]. In addition, the State Council, the Ecological Environment Bureau, the Development and Reform Commission and other departments have all issued a series of policies and guidelines to promote the electrification of vehicles across the country, such as the New Energy Vehicle Industry Development Plan (2021–2035) and the Action Plan for Accelerating the Construction of Charging Infrastructure Along Highways. Although EVs have great advantages in terms of sustainability and environmental protection, the development of EVs in China still faces many challenges. For example, consumers are concerned about the charging and safety performance of EVs [12], and consumers also have “mileage anxiety” and “range anxiety” [13]. Furthermore, the financial incentives associated with the promotion of EVs have placed a significant financial burden on the Chinese government, which is gradually adjusting its financial support policies, such as cutting subsidies, that may affect the promotion of EVs [14]. Thus, in order to decarbonize the transportation sector and achieve sustainable development of the Chinese EVs market, it is important to shift the promotion of EVs from the policy-oriented to the market-oriented approach. Meanwhile, it is also necessary for the government and manufacturers to know the factors that influence the decision-making of Chinese consumers.

EVs not only have the attribute of being the tool to meet consumers’ transportation needs but also have the attribute of being environmentally friendly, such as saving energy and reducing GHG emissions. Thus, logically, the purchase of EVs is both self-interest and altruistic, which means it is influenced by consumers’ self-interest (e.g., saving travelling costs, obtaining financial subsidies) as well moral obligation (e.g., conserving resources, reducing GHG emissions, protecting the ecological environment). However, existing studies on Chinese consumers’ intention to purchase EVs have focused more on analysing consumers’ perceptions of its tool attributes and self-interested factors [5,15,16], and the influence of altruism has not been fully explored. Exploring the psychological factors influencing individuals’ intention to purchase EVs solely from an egoistic perspective, it is easy to overlook the public good nature of pro-environmental behaviour itself. The existence of individual altruism is also influencing pro-environmental behaviours in a way that is not perceived by the subject [17]. Meanwhile, studies have shown that studying pro-environmental behaviours with the theory of rationality or the extended model of pro-environmental theory is conducive to improving the predictive and explanatory power of behaviours [18]. TPB explores the relationship between psychological factors and behaviour from a self-interested perspective and is a classic theory in the field of social psychology [19]. TPB suggests that behavioural intentions are influenced by a combination of subjective norms, attitudes and perceived behavioural control, and it has been applied to research in the fields of sociology, environment and economics. The NAM, proposed by Schwartz (1977) [20], was a major discovery by social psychologists exploring altruism and has been widely used to explain and predict individual pro-social behaviour. The core idea of NAM suggests that people can be guided to engage in altruistic behaviour in a way that stimulates altruistic motivation. That is, the awareness of consequences activates personal norms through ascription of responsibility, which, in turn, lead to the onset of an individual’s pro-social and pro-environmental intentions. Thus, our research aims to integrate TPB and NAM to analyse the effects of altruistic and self-interested factors on Chinese consumers’ intention to purchase EVs, and the results can provide a decision-making reference for policy makers and manufacturers.

In summary, the main contributions of the study are expressed as follows:

• Compared with previous studies from the rational perspective and individual cost-benefit assessment, the uniqueness of this research is that it views consumers’ intention to purchase EVs as a result of the combined effect of individual altruistic and self-interested psychological factors. In other words, the paper attempts to construct a research model to predict Chinese consumers’ intention to purchase EVs by integrating TPB and NAM, which not only compromises the rational and altruistic perspectives but also provides a more comprehensive interpretation of Chinese consumers’ intention to purchase EVs.
• Compared with the existing studies integrating TPB and NAM, this paper further discusses the possible correlation between the two theoretical variables, i.e., it analyses the relationship between awareness of consequences and attitudes, subjective norms and personal norms. Therefore, we can provide ideas for subsequent studies in other areas analysing the adoption of EVs and the field of low-carbon consumption.
• Currently, the Chinese government mainly adopts financial incentives to promote the promotion of EVs, and policies mainly focus on individual self-interested factors. However, with the development of EVs industry, it is worth exploring whether there is a need to formulate policies related to altruistic factors. Thus, in our study, the role of self-interested and altruistic factors on individuals’ intention to purchase EVs is considered, and the conclusions can be used as a reference for policymakers.

The remainder of this study is organised as follows. The literature review and research hypotheses are in Section 2 and Section 3, respectively. Section 4 introduces the research methodology, and Section 5 reports the results of the hypothesis testing. Section 6 discusses the results of the study. Section 7 presents the research implications, and Section 8 reports conclusions and future research directions. In addition, the summary of existing studies about consumers’ intention to purchase EVs is provided in Appendix A.

2. Literature Review

Recent studies on consumers’ intention to purchase EVs are shown in Appendix A. From the recent studies, it can be seen that when exploring the factors influencing the purchase intention of EVs, the studies have mainly focused on the following three aspects:

Firstly, considering EV as an emerging technology, such studies used TAM as the theoretical basis. Most of the findings confirmed the significant influence of perceived ease of use, perceived usefulness and attitudes on consumers’ intention [21,22]. There are also studies that have extended TAM, such as the research by Jaiswal et al. (2021) [22]. Perceived risk and financial incentives were introduced in the study to predict consumer adoption of EVs, and it was found that attitudes, perceived usefulness, perceived ease of use and perceived risk all had direct and indirect effects on intention to adopt EVs.

Secondly, these studies were based on TPB, focusing on the commodity attributes of EVs and analysing the impact of rationality factors on consumers’ purchase intention. In studies applying TPB to analyse consumers’ intention to purchase EVs, consumers’ intention was mainly regarded as a rational choice, and most of them believed that attitudes and subjective norms had a positive effect on intention to purchase EVs [16,23,24,25,26]. However, there was disagreement about whether perceived behavioural control had a significant effect on purchase intention [24,25,26]. For example, Asadi et al. (2021) [1] analysed the factors influencing consumers’ intention to purchase EVs in Malaysia based on 177 questionnaires, and the results showed that perceived behavioural control did not have a significant effect on purchase intention. However, studies by Zhang et al. (2018) [26] and Vafaei-Zadeh et al. (2022) [24] showed that perceived behavioural control was significantly correlated with EV purchase intention. Therefore, this paper raises questions on whether the effect of perceived behavioural control on intention shows different effects in different areas, and it seems necessary to further clarify the relationship between perceived behavioural control and intention. Moreover, there are also preoccupations to expand TPB by introducing other variables. For example, Shanmugavel and Balakrishnan (2023) [23] used TPB as a framework to predict the behavioural intentions of Indian consumers towards EVs by integrating pro-environmental variables such as environmental responsibility, environmental knowledge and environmental concern. Then, Sun et al. (2023) [16] introduced perceived difficulties, personal norms, perceived certainty and environmental concern to expand TPB and comparatively analysed the differences in consumers’ intention to adopt EVs in Hong Kong and Norway. The above analyses reveal that more and more studies have explained or predict consumers’ purchased intention of EVs by expanding TPB.

Finally, studies have pointed out that EVs have environmental attributes, such as resource conservation and the reduction of GHG emissions, and thus the purchase of EVs can be regarded as a pro-environmental behaviour [27]. Meanwhile, scholars have gradually studied the influencing factors of EV adoption from the pro-environmental perspective, and such studies have mainly been based on NAM. For example, He and Zhan (2018) [28] proposed an extended NAM model that included external costs and consumer perceived effectiveness. The results showed that awareness of consequence, ascription of responsibility and perceived effectiveness had a positive effect on personal norms, and the relationship between personal norms and EVs adoption was moderated by external costs. In addition, studies have applied Stimulus–Organism–Response theory (SOR) and self-consistency theory to explore consumers’ purchase intention [29,30].

To summarize, the purchase intention of EVs is the result of a combination of factors involving self-interest and altruistic obligations. However, the existing research lacks a comprehensive consideration of the effects of self-interest and altruistic elements. TPB is a rational choice theory of environmental behaviour [31,32], which views behaviours only from the rational perspective, ignoring the role of altruistic motives in shaping behaviour [28]. According to the TPB, whether or not consumers purchase EVs is the result of a rational trade-off between expected benefits and costs. NAM has been widely used to explain and predict individual pro-social behaviours, the theory suggests that personal norms indicate consumers’ altruistic behaviours and influence their decision-making process towards pro-environmental behaviours. However, NAM only considers the role of moral and pro-social factors in shaping behaviours, arguing that consumers purchase EVs out of moral norms and failing to take into account the effect of rational variables [33]. Although both TPB and NAM have shown good explanatory power in consumer EVs purchase intention, both tend to focus too much on self-interested or altruistic unilateral psychological motives, making it difficult to interpret the whole picture of individual psychological activities [34]. TPB is considered to be highly biased towards self-interest and social responsibility, whereas NAM involves individual environmental motivations and moral obligations, and the integration of TPB and NAM is necessary to resolve the conflicting tensions between TPB and NAM. Currently, the integration of TPB and NAM has been used to predict consumers’ pro-environmental behaviours in the household consumption environment, including recycling, household PM2.5-reduction behaviour, purchase of organic food and residential water conservation behaviour [35,36,37,38]. And lots of studies have shown that a single application of NAM or TPB is insufficient to explain green behavioural intentions [28,39,40]. Therefore, this research argues that integrating TPB and NAM can provide complementary research perspectives for explaining consumers’ EVs purchase intention and help to reveal the full picture of the psychological activities.

3. Research Hypothesis Development

3.1. TPB and the Relationship Among Its Variables

The core principle of TPB holds that behaviour is likely to occur when people perceive that individuals or groups important to them support the behaviour (i.e., subjective norms come into play), have positive attitudes towards the behaviour and believe that they have the ability to adopt the behaviour (i.e., perceived behavioural control produces utility). Subjective norms emphasise the social pressures that individuals feel from the attitudes and behaviours of people or groups that are important to them in their treatment of a behaviour before they commit it. When important people or groups believe that the behaviour should be adopted, the stronger the individual’s intention to act is. Attitudes are the individual’s positive or negative feelings and judgements about a behaviour. When an individual perceives a behaviour more positively, the greater the intention is to adopt that behaviour. Perceived behavioural control is the degree of difficulty and resistance perceived by individuals before the implementation of the behaviour. It measures the individual’s perception of whether or not the behaviour can be accomplished by his or her own volition, and the stronger the volition is, the higher the intention is to perform the behaviour. In this paper, EV purchase intention refers to a consumer’s determination of the likelihood that he or she will purchase an EV, reflecting the consumer’s inclination [19].

As can be seen from the definition, subjective norms emphasise the fact that consumers view their own purchase of EVs from the perspective of those most important to them. When the important people or groups around them believe that they should buy EVs, the stronger the consumers’ intention is to buy. Meanwhile, it has been shown that subjective norms can stimulate purchase intention; for example, Wang et al. (2016) argued that subjective norms can stimulate consumers’ purchase intention for hybrid electric vehicles [41]. Sheng et al. (2019) confirmed that there was a positive effect of subjective norms on Chinese consumers’ green purchase intention [42]. Therefore, this paper proposes the hypothesis:

H1. 

Subjective norms have a positive effect on purchase intention.

In this study, attitudes are considered as the overall positive assessment of consumers’ intention to purchase EVs [26]. According to TPB, attitudes have a positive effect on consumers’ behavioural intentions [19]. Numerous studies have shown that consumers’ attitudes and behavioural intentions are positively correlated. For instance, Wang et al. (2021) [43] showed that attitude was an important predictor of consumers’ intention to purchase battery electric vehicles, and Li and Shao (2023) confirmed that attitudes could significantly stimulate consumers’ intention to purchase eco-friendly clothing [44]. Therefore, this research argues that when consumers’ overall appraisal of EVs is more positive, their intention to purchase EVs is stronger. In summary, the following hypothesis is proposed:

H2. 

Attitudes have a positive effect on purchase intention.

This study argues that perceived behavioural control emphasises consumers’ perceptions of the ease or difficulty of using EVs (Ajzen, 1991) [19]. It has been shown that perceived behavioural control can stimulate individuals’ behavioural intention. For example, Zhao’s (2020) [45] study confirmed that perceived behavioural control could stimulate consumers’ intention to purchase energy-saving home appliances, and the study of Le and Nguyen’s (2022) [37] showed that perceived behavioural control had a positive effect on consumers’ intention to purchase organic food. This study argues that perceived behavioural control emphasises the difficulties consumers feel when purchasing and using EVs. When consumers’ ability to control difficulties is stronger, the intention to purchase EVs is higher. Thus, this research proposes the following hypothesis:

H3. 

Perceived behavioural control has a positive effect on purchase intention.

3.2. NAM and the Relationship Among Its Variables

The NAM was proposed by Schwartz (1977) [20] as a theory to explain and predict individuals’ pro-environmental behaviour and consists of three core variables: personal norms, ascription of responsibility and awareness of consequences. Personal norms are an individual’s sense of moral obligation to adopt a behaviour [46]. Ascription of responsibility refers to an individual’s sense of responsibility for the negative consequences of not adopting a behaviour. Awareness of consequence highlights an individual’s perception of the positive or negative consequences of an action [47]. Personal norms are activated by the conditions of ascription of responsibility and awareness of consequences, which act directly on the individual’s pro-social and pro-environmental intentions. Meanwhile, awareness of consequences influences personal norms indirectly through ascription of responsibility [48].

In this study, awareness of consequence emphasises consumers’ awareness of the negative results of purchasing fuel vehicles, such as GHG emissions and damage to the ecological environment. Ascription of responsibility refer to consumers’ belief that they should take responsibility for the negative consequences caused using fuel vehicles. Previous studies have shown that awareness of consequence can positively affect ascription of responsibility in different pro-social behaviours. For example, Lv et al. (2016) [49] concluded that an individual’s awareness of consequences of non-energy-efficient behaviours has a positive effect on the ascription of responsibility. Therefore, our research argues that when consumers perceive the negative consequences caused by fuel vehicles more strongly, the ascription of responsibility is stronger. The following hypothesis is proposed:

H4. 

Awareness of consequence has a positive effect on ascription of responsibility.

Consumer purchase of EVs is seen as a pro-environmental behaviour, and personal norms relate to consumers’ sense of moral obligation towards the purchase of EVs. This study suggests that when consumers’ perceptions of the environmental problems caused by the use of fuel vehicles are stronger, the more personal norms can be activated. Meanwhile, Shen et al. (2020) [34] showed that the awareness of consequences is the main variable that activates personal norms of household waste sorting among farmers. In summary, the following hypotheses are proposed:

H5. 

Awareness of consequence has a positive effect on personal norms.

It has been shown that ascription of responsibility positively affects personal norms [34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49]. If individuals are aware of the negative consequences of not adopting pro-environmental behaviours, they will feel more responsible for this behaviour and become morally committed to implementing it. Consumers’ ascription of responsibility for the negative outcomes of using fuel vehicles will inspire a sense of moral obligation to conserve resources. Thus, the following hypothesis is proposed:

H6. 

Ascription of responsibility has a positive effect on personal norms.

According to the NAM, individuals will be led to adopt pro-environmental behaviour when their personal norms are activated [20]. Prior research has identified personal norms as an important factor in motivating willingness to engage in pro-environmental behaviours. For example, Wang and Zhang (2017) [17] showed that personal norms positively influenced urban residents’ behavioural intentions to participate in environmental governance. Li and Shao’s (2024) findings confirmed that personal norms could stimulate consumers’ intention to purchase eco-friendly clothing [44]. Zhao’s (2020) study pointed out that there was a positive effect of personal norms on consumers’ intention to purchase energy-saving home appliances [45]. Meanwhile, Le and Nguyen’s (2022) study confirmed that personal norms significantly influenced consumers’ intention to purchase organic food [37]. In our research, personal norms are described as an individual’s sense of moral obligation to purchase EVs. After consumers form personal norms, the sense of moral obligation to purchase EVs is activated, which leads to consumers’ intention to purchase. Therefore, the following hypothesis is proposed:

H7. 

Personal norms have a positive effect on purchase intention.

3.3. The Relationship between TPB and the NAM

TPB and the NAM explain or predict individual behaviours based on different perspectives. TPB takes self-interest as a theoretical starting point and considers self-interest maximisation as the basic principle of individual behavioural decisions. The NAM, on the other hand, emphasises the influence of altruism in individual behavioural decision-making and argues that an individual’s sense of moral obligation is the main driver of pro-environmental behaviour. Purchasing EVs is the result of the influence of multiple factors, including altruistic obligations and self-interest [50], and it is difficult to fully understand consumer purchasing decisions by analysing EVs adoption only from an altruistic or self-interested perspective. Therefore, it is necessary to integrate TPB and NAM to gain an in-depth understanding of consumers’ intention to purchase EVs. For example, Wang and Zhang (2017) [17], after conducting a comparative empirical study of TPB and NAM, found that integrating the two theories into the same theoretical framework was conducive to improving the explanatory power of the model for pro-environmental behaviours.

Subjective norms are antecedent variables of personal norms that determine the social correctness of individual behaviours, which, in turn, leads consumers to test the consistency of their behaviours with their sense of moral obligation [17]. Related studies have also confirmed the link between subjective norms and personal norms [16]. This research argues that the more social pressure consumers perceive on the purchase of EVs, the more they can stimulate their personal norms. Therefore, the following hypothesis is proposed:

H8. 

Subjective norms have a positive effect on personal norms.

Awareness of consequence affects consumers’ attitudes towards adopting a behaviour, as Wang and Zhang (2017) [17] confirmed that consequence awareness of participating in environmental governance positively affected the attitudes of urban residents. Therefore, we believe that if the more optimistic consumers expect the results of purchasing EVs, the more positive attitudes consumers hold. In summary, we propose the following hypothesis:

H9. 

Awareness of consequence has a positive effect on attitudes.

Accordingly, the research model of our study was obtained, as shown in Figure 1.

4. Research Methods

4.1. Measurement Development

The study involves seven variables, and the measurement questions for each variable were derived from well-established scales with appropriate modifications. Among them, the measurement items of subjective norms, attitudes, perceived behavioural control and purchase intention were borrowed from the studies of Ajzen (1991) [19], Schniederjans and Starkey (2014) [51] and Huang and Ge (2019) [52]. The measurement items of awareness of consequence, ascription of responsibility and personal norms refer to the studies of Li et al. (2023) [53], Zhang et al. (2013) [54] and Du et al. (2022) [55], respectively. The details are shown in

Table 1. Measurement question items for variables.

Variables	Code	Items
SN	SN1	News media publicity can influence my decision to purchase EVs.
	SN2	Encouragement from friends around me will increase my intention to purchase EVs.
	SN3	The outcry will make me consider buying EVs to reduce environmental pollution.
ATT	ATT1	I think buying EVs will be better for the environment.
	ATT2	I support more state policies to encourage individuals to purchase EVs.
PBC	PBC1	If I wanted to buy an EV, I could afford its current price.
	PBC2	When I decide to use EVs, I can afford to do so, even if the maintenance costs may be slightly higher.
	PBC3	I have the time and access to information about EVs.
AC	AC1	I think buying fuel vehicles accelerates global warming.
	AC2	I think buying fuel vehicles accelerates the shortage of resources.
	AC3	I think buying fuel vehicles pollutes the environment.
AR	AR1	We are responsible for the adverse health effects of fuelled vehicles’ exhaust.
	AR2	We are responsible for the environmental degradation caused by the use of fuel vehicles.
	AR3	We are responsible for the GHG produced by the use of fuel vehicles.
PN	PN1	It is my responsibility to do my part to protect the environment and conserve resources.
	PN2	I will take the initiative to learn about environmental protection.
	PN3	Although my own influence is small, I want to contribute to environmental protection.
	PN4	We should all realise the need to protect the environment; otherwise, our future generations will suffer the consequences.
PI	PI1	When I need to purchase vehicles, I prefer EVs.
	PI2	When I need to replace my vehicle, I prefer to buy an EV.
	PI3	EVs are my first choice for future vehicle purchases.

Note: subjective norms = SN; attitudes towards EVs = ATT; perceived behavioural control = PBC; awareness of consequence = AC; ascription of responsibility = AR; personal norms = PN; purchase intention to EVs = PI.

. All the constructs have a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Before the formal research, due to the changes of the initial maturation scale in terms of language transformation and research context, the pre-testing data needed to be subjected to exploratory factor analysis (EFA) in order to correct the scale entries. The pre-testing questionnaire was distributed through the ‘Questionnaire Star’ website (http://www.sojump.com/). In order to ensure the quality and authenticity of the questionnaire data, two ‘concentration test’ questions were designed to identify those who did not answer the questionnaire attentively. A total of 175 questionnaires were obtained from the pre-testing, and 167 valid questionnaires were obtained after excluding invalid questionnaires, with an effective recovery rate of 94.35%. The data obtained from the pre-testing were subjected to EFA using SPSS v.26.0 through principal component analysis and maximum variance rotation according to the criterion that the root of identity is greater than 1. The results are shown in

Table 2. Results of pre-testing.

Variables	Code	Loading 1	Loading 2	Loading 3	Loading 4	Loading 5	Loading 6	Loading 7
ATT	ATT1	0.766
	ATT2	0.766
SN	SN1		0.653
	SN2		0.773
	SN3		0.764
PBC	PBC1			0.734
	PBC2			0.774
	PBC3			0.617
AC	AC1				0.649
	AC2				0.694
	AC3				0.791
AR	AR1					0.679
	AR2					0.797
	AR3					0.756
PN	PN1						0.708
	PN2						0.824
	PN3						0.832
	PN4						0.707
PI	PI1							0.796
	PI2							0.765
	PI3							0.776

. The KMO value was 0.713, the Bartlett’s test statistic was significant at the 0.001 level, and the variance contribution rate was 76.406%, indicating that the variables had good measurement validity. In addition, the reliability coefficient Cronbach’s α values for each variable were greater than 0.7, and the measurement scales were internally consistent. Therefore, the scales could be used in formal questionnaires.

4.2. Data Collection

The formal questionnaire survey was conducted in January 2023. The formal research adopted the snowball approach to collect data, distributing online questionnaires to the subject group, family members, friends, colleagues and neighbours, and inviting them to pass the questionnaires to their colleagues and friends around them. Finally, 922 questionnaires were obtained, and after excluding invalid questionnaires, 867 valid questionnaires were obtained, and the validity rate of the questionnaires was 94.03%. The scope of the study basically covers 32 provinces, autonomous regions and municipalities in China. The descriptive statistics of the sample are shown in

Table 3. Analysis results of the demographic variables (n = 867).

Demographic Variable	Items	Frequency	Percentage
Gender	Male	445	51.33%
	Female	422	48.67%
Age	18–20	33	3.81%
	21–35	409	47.17%
	36–50	310	35.76%
	51–65	110	12.69%
	>65	5	0.57%
Education Level	High school or below	84	9.69%
	Associate degree	68	7.84%
	Bachelor’s degree	268	30.91%
	Master’s degree or above	447	51.56%
Total household income for the previous year (RMB)	≤30,000	54	6.23%
	30,000–100,000	256	29.53%
	100,000–250,000	331	38.18%
	250,000–500,000	160	18.45%
	>500,000	66	7.61%

.

As we can see from Table 3, 48.67% of the participants were female, and 51.33% were male. As of 2021, 48.81% of China’s population was female, and 51.19% was male [56]. The proportion of males participating in our research was slightly higher than the proportion of males in the total Chinese population, while the proportion of females was lower than the proportion of females in the total population. However, since these differences were small (less than 10 percent), the data could be considered well distributed in terms of gender.

In terms of age distribution, according to the age statistics of National Bureau of Statistics (2022) [56], the proportion of the Chinese population aged 20–34, 35–49 and 50–69 is 19.88%, 21.54% and 27%, in that order. It can be found that the age distribution had a large difference between the sample and the total population. Although Huang and Ge (2019) [52] noted that younger groups were more concerned about EVs. In order to avoid the age distribution affecting the validity of the findings, the one-way ANOVA was conducted to determine if there was statistical evidence that the age groups were significantly different in the means of each variable. The results of the one-way ANOVA indicated that the age groups were not significantly different on any of the variables. Therefore, it could be assumed that age did not bias the results of the hypothesis test.

The education level of the sample was higher than that of the total Chinese population. Although the group with higher education level is more concerned about EVs [52], the one-way ANOVA was still required. The results showed that there was no significant difference between the education groups on any of the variables. Therefore, the higher education level of the sample did not seem to lead to bias when compared to the entire Chinese population.

According to the National Bureau of Statistics, in 2021, the national median disposable income per capita was about RMB 30,000, which means that half of the households in China have an income below RMB 30,000 [56]. In our study, the percentage of the same income group was 6.23, so the one-way ANOVA was required. The results show that there was no significant difference in the means between the income groups on the independent and dependent variables. Therefore, differences in income levels between the sample and the overall population did not seem to affect the results of the study.

4.3. Statistical Analysis

4.3.1. Common Method Bias

In order to ensure the correctness and scientificity of the sample data, it was necessary to test whether the sample data conformed to the normal distribution to avoid the influence of the non-normal distribution of the data on the results of the hypothesis test of SEM. We used SPSS v.26.0 to test the skewness and kurtosis of the sample data. After analysis, the absolute values of the skewness coefficient of the sample data ranged from 0.288 to 1.911, none of which was greater than 3, and the absolute values of the kurtosis coefficient ranged from 0.030 to 4.808, none of which was greater than 8. This indicates that the sample data conformed to the normal distribution and could be used for hypothesis testing in SEM [57].

Additionally, the sample data were all from online questionnaire platforms, and self-reported responses were used, which may have led to common method variance in the test results and affect the validity of the findings. Therefore, the common method variance of the sample data needed to be controlled and tested. Common method variance was procedurally controlled during the data collection process by taking measures of anonymity, selection of questions and partial attention to the test items. For testing, the common method variance of the sample data, Harman’s one-way test [58], was adopted, and exploratory factor analysis was conducted on all latent variables. The results showed that the exploratory factor analysis extracted a total of 5 factors (greater than 1) with eigenroots greater than 1, and the largest factor variance explained was 40.077% (less than 50%) [59], thus indicating that the data did not have common method bias. And there was no significant improvement in the model parameters after introducing common method bias as a latent variable into SEM, indicating that common method bias was not addressed in our research. Meanwhile, according to the results of confirmatory factor analysis (CFA), the data fitting effect of the single-factor model (χ²/df = 28.624 > 3, CFI = 0.553 < 0.9, TLI = 0.504 < 0.9, RMSEA = 0.179 > 0.05, SRMR = 0.1175 > 0.05) was significantly less satisfactory than that of the seven-factor model. Therefore, there was no significant common method bias in the measurements.

4.3.2. Reliability and Validity

The reliability and validity tests were conducted using SPSS v.26.0 and Amos v.25.0, and the results are presented in

Table 4. Analysis results of convergent validity and reliability.

Latent Variables	Items	Loading	Cronbach’ s α	CR	KMO	AVE
ATT	ATT1	0.830	0.810	0.811	0.500	0.682
	ATT2	0.822
PBC	PBC1	0.775	0.739	0.750	0.650	0.507
	PBC2	0.794
	PBC3	0.537
SN	SN1	0.790	0.844	0.844	0.729	0.644
	SN2	0.799
	SN3	0.818
AC	AC1	0.874	0.927	0.928	0.758	0.811
	AC2	0.894
	AC3	0.933
AR	AR1	0.670	0.835	0.845	0.675	0.647
	AR2	0.879
	AR3	0.849
PN	PN1	0.716	0.883	0.888	0.796	0.668
	PN2	0.891
	PN3	0.914
	PN4	0.729
PI	PI1	0.857	0.912	0.887	0.759	0.724
	PI2	0.849
	PI3	0.847

and

Table 5. Analysis results of discriminant validity.

	Mean	ATT	PBC	SN	AC	AR	PN	PI
ATT	3.893	0.826
PBC	3.421	0.395 ***	0.712
SN	4.399	0.484 ***	0.308 ***	0.802
AC	4.011	0.637 ***	0.295 ***	0.687 ***	0.901
AR	4.012	0.748 ***	0.379 ***	0.553 ***	0.645 ***	0.805
PN	3.234	0.627 ***	0.392 ***	0.431 ***	0.468 ***	0.513 ***	0.817
PI	3.019	0.476 ***	0.361 ***	0.373 ***	0.329 ***	0.347 ***	0.565 ***	0.851

Note: *** p < 0.001.

. The results revealed that the factor loadings of all items were higher than the threshold of 0.5, indicating that the questionnaire exhibited good construct validity. The results showed that the Cronbach’s α value of each variable was higher than 0.7, indicating that the scale had high reliability. Since attitude is a 2-item scale, the Spearman–Brown coefficient was also needed as a reliability indicator, and the results show that the Spearman–Brown coefficient for attitude was 0.811, which is higher than the recommended standard of 0.70, indicating that the reliability was good.

In order to test whether the scale items could truly and accurately reflect the measurement objectives, model fit adaptability was judged by CFA. In our study, the CFA was conducted using Amos v.25.0, and the model and path coefficients are shown in Figure 2. The model fit test is shown in

Table 6. Analysis results of the validation factor analysis.

Fit Index	Goodness-of-Fit Index	Evaluation Criteria	Values	Evaluation Results
Absolute fit index	χ2/df	<3, good fit; <5, reasonable fit	2.951	good fit
	GFI	>0.9, good fit	0.945	good fit
	RMSEA	<0.05, good fit; <0.08, reasonable fit	0.047	good fit
Relative fit index	NFI	>0.9, good fit	0.958	good fit
	TLI	>0.9, good fit	0.965	good fit
	CFI	>0.9, good fit	0.972	good fit
	IFI	>0.9, good fit	0.972	good fit

. From the test results, it can be seen that the CFA model had a good fit. Then, convergent validity and discriminant validity analyses were performed. The results of convergent validity analysis are shown in Table 4, the standardised factor loading coefficients of each measurement item were greater than 0.5, the average variance extracted (AVE) of each dimension was greater than 0.5 and the composite reliability (CR) was greater than 0.7, which indicates that each dimension has good convergent validity. The results of the discriminant validity analysis are shown in Table 5, and the standardised correlation coefficients between the two dimensions were smaller than the square root of the corresponding AVE of the dimensions, indicating that each dimension had good discriminant validity. In summary, the scale had sufficient reliability and validity.

5. Results

5.1. Hypotheses Testing Results

The study applied Amos v.25.0 to test the SEM and the overall fit indices are shown in

Table 7. Fitting indices of the SEM.

Fit Index	Goodness-of-Fit Index	Evaluation Criteria	Values	Evaluation Results
Absolute fit index	χ2/df	<3, good fit; <5, reasonable fit	4.734	reasonable fit
	GFI	>0.9, good fit	0.913	good fit
	RMSEA	<0.05, good fit; <0.08, reasonable fit	0.066	reasonable fit
Relative fit index	NFI	>0.9, good fit	0.930	good fit
	TLI	>0.9, good fit	0.933	good fit
	CFI	>0.9, good fit	0.943	good fit
	IFI	>0.9, good fit	0.944	good fit

, which indicates a good fit for the SEM.

The results of hypothesis testing are shown in

Table 8. Results of the hypothesis test.

Path				Estimate	S.E.	C.R.	p	Results
H1	SN	→	PI	0.086	0.057	2.08	*	Supported
H2	ATT	→	PI	0.136	0.045	2.997	**	Supported
H3	PBC	→	PI	0.138	0.04	3.533	***	Supported
H4	AC	→	AR	0.672	0.04	18.707	***	Supported
H5	AC	→	PN	0.191	0.066	3.298	***	Supported
H6	AR	→	PN	0.303	0.05	6.311	***	Supported
H7	PN	→	PI	0.416	0.039	9.93	***	Supported
H8	SN	→	PN	0.141	0.076	2.745	**	Supported
H9	AC	→	ATT	0.669	0.04	17.822	***	Supported

Note: *** p < 0.001, ** p < 0.01, * p < 0.05.

, and the standardised path coefficients of SEM are shown in Figure 3. There was a significant positive effect of subjective norms on purchase intention, with a regression coefficient of 0.086 and significant at the 5% level. It indicates that when important people or groups around the consumer believed that EVs should be purchased, the stronger the consumer’s intention was to purchase EVs, and H1 was supported. Attitudes had a positive effect on the purchase intention of EVs and were significant at the 1% level, with a regression coefficient of 0.136. It can be seen that when consumer’s attitudes towards EVs were positive, the greater the purchase intention was, so H2 was confirmed. Perceived behavioural control could stimulate the intention to purchase EVs, the regression coefficient was 0.138 and significant was at the 0.1% level, which shows that when consumers believed that they had the ability to use EVs, the greater the intention was to purchase EVs, so H3 was supported. Awareness of consequence could significantly stimulate consumers’ ascription of responsibility and personal norms, and both were significant at the 0.1% level, with regression coefficients of 0.672 and 0.191, respectively, suggesting that consumers’ perceptions of the adverse results of using fuel vehicles could significantly stimulate consumers’ ascription of responsibility and personal norms. Therefore, H4 and H5 were valid. There was a significant positive effect of ascription of responsibility on personal norms with a regression coefficient of 0.303 and significance at the 0.1% level. It can be seen that when consumers felt more responsible for purchasing EVs, the more it could stimulate consumers’ sense of environmental protection moral obligation, so H6 was valid. There was also a significant positive effect of personal norms on consumers’ intention to purchase EVs, with a regression coefficient of 0.416, which was significant at the 0.1% level. It means that consumers’ intention increased when they had a stronger sense of moral obligation to purchase EVs, hence H7 was confirmed. Additionally, subjective norms positively influenced personal norms, with a regression coefficient of 0.141 and significant at the 1% level. It suggests that when consumers perceived more social pressure to purchase EVs, the more they were able to motivate their sense of moral obligation, and H8 was supported. Awareness of consequence significantly motivated consumer attitudes towards EVs and was significant at the 0.1% level with a correlation coefficient of 0.669, i.e., consumers’ perceptions of the negative consequences of using fuel vehicles affected their attitudes towards EVs, so H9 was valid.

5.2. Results of the Mediation Effects Test

The common methods of mediation effects tests are mainly causal, Sobel and Bootstrap. Among them, Bootstrap is not only the most effective but also does not have any requirement on sample distribution, which is regarded as the preferred method for mediation effect tests [60]. Therefore, the Bootstrap of Amos v.25.0 was used in our research for the mediation test. Setting the number of replications K = 5000 and confidence interval CI% = 95%, the results of the mediation effects test are shown in

Table 9. Results of the mediation effects.

Path	Point Estimates	SE	Bias-Corrected 95%CI			Percentile 95%CI
			Lower	Upper	p	Lower	Upper	p
AC → AR → PN	0.203 ***	0.044	0.123	0.292	0	0.122	0.291	0
AC → PN → PI	0.080 *	0.032	0.019	0.144	0.014	0.017	0.142	0.017
AC → AR → PN → PI	0.085 ***	0.021	0.05	0.132	0	0.048	0.13	0
AC → ATT → PI	0.091 *	0.035	0.024	0.162	0.012	0.022	0.159	0.015
SN → PN → PI	0.059 **	0.025	0.015	0.113	0.007	0.012	0.109	0.01

Note: *** p < 0.001, ** p < 0.01, * p < 0.05.

. As can be seen from Table 8, the confidence intervals for all mediated paths did not include 0, indicating that all mediated paths in the SEM held.

6. Discussion

The results of SEM indicate that both self-interested and altruistic psychological attributes had a significant driving effect on Chinese consumers’ intention to purchase EVs, validating the rationality of SEM. It is worth noting that although subjective norms (β = 0.086), attitudes (β = 0.136) and perceived behavioural control (β = 0.138) all had a significant effect on purchase intention, which verifies the validity of the TPB in explaining the purchase intention of Chinese consumers of EVs, the standardized path coefficients between the variables were relatively small, which differs from that of existing studies. The reason may be that with the development of the EVs industry, the gap between EVs and fuel vehicles in terms of the vehicle’s tool attributes to meet consumers’ transport needs has gradually narrowed, and the environmental attributes of EVs are the main factors driving consumers’ purchases.

There have been controversies about the relationship between subjective norms and consumers’ intention to purchase EVs, e.g., Shanmugavel and Balakrishnan (2023) [23] argued that subjective norms could not motivate Indian consumers’ intention to purchase EVs. However, our findings indicate that subjective norms could significantly influence Chinese consumers’ purchase intention for EVs, consistent with the findings of Zhang et al. (2018) [26], Asadi et al. (2021) [1] and Ackaah et al. (2022) [25]. The reason may lie in the fact that the relationship between subjective norms and consumers’ purchase intention varies across cultures. In China, Confucius’s culture has been influencing the thoughts and actions of people, and this influence can be reflected in collectivist values. In this group-oriented cultural atmosphere, people are more encouraged to obey the organisation than to pursue individual goals. As a result, consumer decision-making is influenced by groups such as family, neighbours, friends, co-workers and even society as a whole [61,62]. And it has been shown that collective culture affects the formation of individual subjective norms, which, in turn, affects individuals’ intention to consume low carbon [63]. This also suggests the need to pay attention to the cultural background of consumers when exploring the intention to purchase EVs [64].

The results show that there was a significant positive effect of attitudes on the intention to purchase EVs, which is largely consistent with existing studies, and that attitudes were stimulated by the awareness of consequences. In our study, awareness of consequences was based on the assumption that consumers perceived the negative impacts of using fuel vehicles, such as polluting the environment and accelerating resource scarcity. It suggests that consumers’ perceptions of the negative impacts associated with fuel vehicles will increase consumers’ positive attitudes towards EVs and the greater the likelihood that they will ultimately purchase EVs. In addition, there was a significant effect of awareness of consequences on ascription of responsibility, and it has been shown that the two do have a significant correlation [1,28]. It suggests that when consumers’ knowledge of the environmental disadvantages of fuel vehicles is deeper, consumers’ sense of responsibility for the negative consequences of not using EVs is stronger. Meanwhile, it was found that Chinese consumers’ intention to purchase EVs was stronger when their perceived behavioural control was higher, which is consistent with the findings of Zhang et al. (2018) [40] and Vafaei-Zadeh et al. (2022) [24].

The SEM results indicate that personal norms played the greatest role in influencing the intention to purchase EVs, which is true, as Stern et al. (1999) [65] stated that personal norms are considered as important psychosocial factors that can positively influence individuals’ intention in environmentally friendly manners. It suggests that consumers’ sense of environmental obligation towards the purchase of EVs is the main factor influencing their intention to do so, and therefore, the in-depth understanding of the environmental attributes of energy efficient products is important to stimulate environmentally friendly behaviour. At the same time, personal norms were also stimulated by the awareness of consequences and ascription of responsibility, consistent with existing research [1,28]. The results of our paper illustrate that the deeper consumers’ perception was of the environmental problems caused by fuel vehicles, the more they were stimulated to feel a sense of environmental obligation towards purchasing EVs. In addition, individuals had a sense of responsibility for the outcome of their behaviour, and when individuals recognised the positive outcomes of adopting environmentally friendly behaviours, this led to a sense of moral obligation to perform the behaviour [66], i.e., consumers’ awareness of the environmental problems associated with the use of fuel vehicles could motivate individuals to develop a sense of moral obligation to be environmentally friendly. In addition, subjective norms had a positive effect on personal norms, suggesting that social pressures from the outside, as well as evaluations from others, internalise consumers’ sense of morality regarding the purchase of EVs and motivate purchase intention.

7. Implications

7.1. Theoretical Implications

The main findings contributions are as follows:

• Personal norms, perceived behavioural control, attitudes and subjective norms had a significant positive effect on consumers’ EVs purchase intention, with personal norms playing the largest role. It indicates that altruistic factors played a crucial role in motivating consumers’ EV purchase intention. Meanwhile, awareness of consequences, ascription of responsibility and subjective norms were positive predictors of persona norms.
• Awareness of consequence, ascription of responsibility and subjective norms had indirect effects on EVs purchase intention, where the indirect effect of awareness of consequence was realised through the mediating paths of ascription of responsibility, personal norms and attitudes; both ascription of responsibility and subjective norms had an indirect effect on EV purchase intention through personal norms.
• Between TPB and the NAM, subjective norms could stimulate consumers’ personal norms, and awareness of consequence could significantly increase Chinese consumers’ positive attitudes towards EVs.
• The results of this study confirmed that subjective norms could stimulate Chinese consumers’ intention to purchase EVs, clarifying the relationship between subjective norms and purchase intention. The reason for the discrepancy between this finding and existing studies may be that cultural background affects the formation of individual subjective norms. Therefore, subsequent research on EVs adoption should consider the cultural background of consumers.

7.2. Practical Implications

The conclusion can provide a reference for the development of EVs related promotion policies and marketing strategies. The results of data analyses show that self-interested factors had a smaller role in driving the purchase intention of EVs, and altruistic factors had a larger influence. Therefore, policymakers and manufacturers should pay attention to the environmental advantages of EVs.

Firstly, consumers with higher personal norms showed higher purchase intention of EVs, so the government should adopt various ways to stimulate consumers’ sense of environmental obligation. For example, the government can strengthen the publicity and education activities on the negative effects of carbon emissions, advocate low-carbon consumption concepts and methods through official media, improve consumers’ awareness and recognition of the environmental attributes of EVs and encourage consumers to purchase EVs as much as possible, so as to stimulate consumers’ responsibility and obligation to protect the environment through the purchase of EVs.

Second, awareness of consequence could play a role in consumers’ purchase intention of EVs through ascription of responsibility, personal norms and attitudes. Thus, governments and manufacturers should publicise the negative consequences of using fuel vehicles, as well as the positive impacts of EVs. The government can also use social media to educate the public about global warming, environmental pollution and oil risks that may result from the adoption of fuel vehicles. Manufacturers should actively promote the environmental outcomes that driving with EVs can bring, such as the amount of fossil fuels that can be saved and the reduction of GHG emissions, when developing marketing strategies for EVs. Consumers’ awareness of the negative consequences of using fuel vehicles and the perception of the positive consequences of EVs will help them realise their contribution to solving environmental problems. This results in a responsibility and obligation for consumers to purchase EVs and increases positive attitudes towards EVs.

Thirdly, subjective norms not only directly drove consumers’ purchase intention of EVs but also influenced consumers’ purchase intention through personal norms. Therefore, the government should formulate effective guidelines and norms for low-carbon consumption to create a favourable atmosphere for low-carbon consumption in the whole society, which will strengthen the individual’s moral obligation towards low-carbon consumption and stimulate the individual’s willingness to consume low-carbon products.

Fourth, attitudes had a direct impact on EVs purchase intention. Therefore, when developing marketing strategies for EVs, enterprises should pay attention to establishing a positive corporate environmental image to promote consumers to form a positive attitude towards EVs and then stimulate consumers’ purchase intention.

Lastly, perceived behavioural control had a positive effect on consumers’ intention to purchase EVs. Thus, companies can ensure that consumers can easily identify and purchase EVs that meet their needs by innovating EVs styles and expanding publicity channels.

8. Conclusions

In order to decarbonise China’s transport sector and mitigate global warming, attention needs to be paid to the detailed information that influences consumers’ decisions to purchase EVs. Firstly, purchasing EVs not only involves consumers’ individual interests but also has environmental advantages such as saving resources and reducing GHG emissions. Therefore, our study integrated TPB and the NAM to construct the research model of Chinese consumers’ intention to purchase EVs, taking into account both self-interest and altruistic factors. Secondly, based on 867 questionnaires from Chinese consumers, SEM was applied to analyse the results, which showed that both altruistic and self-interested factors can stimulate consumers’ intention to purchase EVs, and the relationship between perceived behavioural control, subjective norms and purchase intention was clarified, with personal norms having the greatest impact on EVs purchase intention. Finally, awareness of consequence, ascription of responsibility and subjective norms can stimulate personal norms, and awareness of consequence can influence consumers’ attitudes towards EVs.

Although our research questions are of practical significance and the conclusions are of theoretical value and practical usefulness, the influence of time, data sources and other factors has led to deficiencies in the following aspects, which need to be improved in future research. Firstly, in terms of research content, although the factors affecting consumers’ intention to purchase EVs were all derived from TPB and the NAM, which have theoretical basis, they do not reflect the specificity of the theoretical model to explain the purchase intention of EVs. So, it can be improved by applying the rooting method and QCA in the future. Second, we used EVs as the research object. However, due to the differences in performance, appearance and price of EVs sold by different manufacturers, this study did not select a representative EVs of a certain category as the research object, which may have led to the authenticity of the respondents’ answers and the applicability of the conclusions being affected. In the future, surveys should be conducted for a certain type of EVs to improve the applicability of conclusions. Once again, our results suggest that it is necessary to pay attention to the cultural background of consumers when exploring their purchase intention of EVs. Follow-up studies could conduct cross-sectional research in multiple regions to comparatively analyse the differences in the effects of different factors on consumers’ purchase intention, which could help to inform policymakers in different regions (e.g., the study by Sun et al., 2022 [16]). Finally, in terms of the research sample, due to the limitation of objective factors, this study only used online promotion to distribute questionnaires and collect data. The attention focus test questions were set to ensure the data quality as much as possible and a sufficient number of valid questionnaires were collected. However, since the data were not collected through offline field research, it is difficult to identify whether the respondents have real spending power. The follow-up studies can adopt the form of mixed online and offline research to improve the scientificity and validity of the sample.